Selfish mining is a strategic deviation from the honest Bitcoin mining protocol in which a mining pool, upon finding a new block, does not immediately broadcast it to the network. Instead, the pool secretly continues mining on its private chain, only publishing blocks strategically to cause honest miners to waste computational effort on stale blocks. The attack was formally described by Ittay Eyal and Emin Gün Sirer in a 2013 paper, causing significant debate because it demonstrated that Bitcoin’s incentives were less perfectly aligned than previously assumed — and that a pool as small as ~25-33% hashrate could theoretically be profitable using the strategy.
How Selfish Mining Works
The following sections cover this in detail.
Normal (Honest) Mining:
- All other miners receive it, verify it, and start mining the next block on top of it
- Everyone is on the same chain; everyone earns proportional to their hashrate
Selfish Mining Strategy:
- Pool continues mining privately, working toward a second block (now secretly 2 blocks ahead)
- If an honest miner finds a block and broadcasts it, the selfish pool immediately releases its private block
A chain race begins (both blocks at same height)
The selfish pool’s block may be adopted first if it has connectivity advantage
If the selfish pool has already found block 2, it releases block 2 immediately — its chain definitively wins, honest block is orphaned - If the pool successfully builds a 2-block lead: releases both blocks, orphaning all work done by honest miners on the competing chain
The payoff: Honest miners’ work gets orphaned. The selfish pool earns block rewards that would otherwise have gone to honest miners.
The 25% Threshold
Eyal and Sirer’s original analysis showed that a selfish miner with ≥25-33% hashrate (depending on the attacker’s network connectivity) could earn more than their proportional share of block rewards by employing this strategy. Below 25%, the strategy is unprofitable (expected earnings are less than honest mining).
Critically: This means a selfish mining pool with 25% hashrate could earn more rewards per unit of hashrate than honest miners — creating an incentive for rational profit-maximizing miners to join the selfish pool, potentially leading to centralization.
Why Hasn’t It Happened?
If selfish mining is theoretically profitable, why haven’t major pools used it?
Game theory complications:
- If other rational pools detected selfish mining, they would also switch strategies, negating the advantage
- Large pools have reputational stakes that theoretical hashrate gains don’t compensate for
- Pool operators are known entities — coordinating a long-term attack would be discoverable
Detection:
- Selfish mining causes a detectable increase in stale/orphan block rates on-chain
- Blockchain analysts would likely detect a sustained selfish mining campaign
Countermeasures:
- Bitcoin nodes can be modified to use unforgeable timestamps or implement “tie-breaking” rules that negate the attacker’s connectivity advantage
- BIP-102 and subsequent proposals addressed some aspects
Practical costs:
- Building and maintaining a large pool creates legal exposure
- Shorter mining hardware ROI if the pool is eventually identified and users leave
Relation to 51% Attack
Selfish mining is related to but distinct from a 51% attack:
| Selfish Mining | 51% Attack | |
|---|---|---|
| Hashrate needed | ~25-33%+ | >50% |
| Goal | Disproportionate block rewards | Double-spend, chain reorganization |
| Effect | Reduces honest miner revenue | Can reverse confirmed transactions |
| Detectability | Moderate (orphan rate increase) | Obvious (requires sustained chain reorg) |
Selfish mining doesn’t enable double-spends. It is purely an economic attack on mining revenue distribution.
Common Misconceptions
“Selfish mining requires a majority of hashrate”
The point of the Eyal/Sirer paper is that selfish mining is theoretically profitable at below 50% — specifically as low as 25-33% hashrate. This was the controversial finding that challenged Bitcoin’s assumed security model.
“Selfish mining has already been used”
There is no confirmed documented case of deliberate selfish mining on Bitcoin. Orphan blocks and temporary chain divergences occur naturally; sustained selfish mining campaigns leave a detectable statistical signature that has not been observed.
Social Media Sentiment
Selfish mining is a topic primarily discussed among technical Bitcoin researchers, academics, and mining industry observers. Eyal and Sirer’s 2013 paper was controversial when published — some Bitcoiners dismissed it as theoretical while others found it alarming. The paper influenced improvements to Bitcoin Core’s peer behavior and spawned significant academic follow-up research. Mining pools occasionally get accused of selfish mining during periods of high orphan rates, though none have been conclusively proven. The topic resurfaces periodically as Bitcoin mining centralization debates continue.
Last updated: 2026-04
Related Terms
Sources
- Eyal, I., & Sirer, E. G. (2014). Majority Is Not Enough: Bitcoin Mining Is Vulnerable. Financial Cryptography and Data Security.
- Sapirshtein, A., Sompolinsky, Y., & Zohar, A. (2016). Optimal Selfish Mining Strategies in Bitcoin. Financial Cryptography and Data Security.
- Nayak, K., Kumar, S., Miller, A., & Shi, E. (2016). Stubborn Mining: Generalizing Selfish Mining and Combining with an Eclipse Attack. IEEE EuroS&P.